OpenAI Agents SDK BMS Telemetry Maintenance Workflows with v0
Develop an autonomous facility agent using the OpenAI Agents SDK and Vercel v0 that converts BMS IoT telemetry streams into prioritized maintenance work orders. The solution optimizes HVAC dispatch schedules while cutting false-alarm rates below 5%.
What you are building
The core problem, expected build, and operating context for this challenge.
Uses OpenAI Agents SDK tool functions coupled with Vercel v0 UI elements to automate Building Management System (BMS) fault detection and work order generation.
How work is evaluated
Evaluates fault isolation precision and dispatch ticket accuracy from IoT telemetry.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
False Alarm Prevention Test
Verifies normal telemetry readings do not trigger critical work orders
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Fault Classification Precision
Accuracy of identified HVAC diagnostic codes • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Implement OpenAI Agents SDK logic using structured functions and tool calls
Generate reactive maintenance UI components using Vercel v0 design workflows
Detect anomalies in real-time HVAC pressure and temperature telemetry streams
Prioritize dispatch schedules based on energy loss and tenant impact
Reference links and supporting material
500 time-series telemetry events from commercial building Air Handling Units and chillers.
Challenge setup
Evaluates fault isolation precision and dispatch ticket accuracy from IoT telemetry.
Challenge input
JSON containing timestamp, sensor_id, temp_f, duct_pressure_psi, and chiller_status
OpenAI Agents SDK
Provides function calling and state management for autonomous workflows.
Vercel v0
Generates maintenance ticket and operations dashboard components.
Evaluated output
JSON with fault_detected (bool), issue_type, urgency_level, and recommended_action
- Verifies normal telemetry readings do not trigger critical work orders
- Accuracy of identified HVAC diagnostic codes • target: 0.95 • range: 0-1
- Benchmark: IoT-FaultDiagnostics-v1
- Fault Classification Precision target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- OpenAI · Required
- Vercel v0 · Optional
- Larridin · Optional
Evaluation contract
- False Alarm Prevention Test · Weight 1
- Fault Classification Precision · Weight 1
Metrics
- Fault Classification Precision · Target 0.95
Benchmarks
- IoT-FaultDiagnostics-v1
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing timestamp, sensor_id, temp_f, duct_pressure_psi, and chiller_status
Output: JSON with fault_detected (bool), issue_type, urgency_level, and recommended_action
Execution harness: Python
Evaluation tools
- OpenAI Agents SDK · Agent Framework
- Vercel v0 · UI Engineering Tool
Recipe freeze marker
No recipe freeze time is recorded.